Safe to use an active backplane?

[StephenFry]

I really enjoy building FreeNAS systems, and have put together quite a few machines, for for myself, for fun, and for friends/fam.
After initial trial-and-error, all builds are now similar: SuperMicro board, one or two Dell or IBM controllers as LSI 9211-8i in IT mode, and a couple of hard drives.
All of these builds were done using re-purposed, or home-built common pc-cases.

For a new build, I'd like to use the same internals as I always do, but in a 16-bay proper 19-inch server chassis. With only theoretical knowledge of this stuff, I needed to get "hands-on" and picked up an affordable second-hand chassis. So far, I've learned a lot and it has been a lot of fun.

More importantly: all initial tests are good. Although there were some weird things about drives not reporting SMART data or not getting recognized with their respective brand names, just as generic "ATA" drive. They did work fine, though.

...or did they?

And there is my question: this chassis uses AIC TW-000-52072-AR R.A1 backplanes, equipped with a SEP (Storage Enclosure Processor) LC4032Y-75TN-101 SGPIO.
Now... I don't really have a clue what that means, but the only info I was able to find in my research, was that this is a so-called 'ACTIVE' backplane - as opposed to a PASSIVE one.

Is this backplane safe to use in a FreeNAS ZFS system?

The only user report on these backplanes that I could find was of someone finding out their Adaptec controller didn't play nice with this specific model, because it is 'active'.

It clearly works for a lot of other people, but they aren't using FreeNAS. Promise SuperTrak controllers and various Areca RAID controllers are reported to work fine.

The last thing I want, is something injecting extra bits and bobs in my data. Very early on in my NAS career, I remember the fake-RAID Sil3032 (?) cards, which -even in IT mode- used to write a bit of information on the last 512 bytes of each drive they encountered. Not something ZFS enjoys...

I don't need the actual functionality, like hotswapping, of a backplane, it's just a convenient way to connect the drives.
If there is even the slightest bit of doubt, I'm pulling the entire backplane and will connect the drives directly.

 

[Jailer]

Is this backplane safe to use in a FreeNAS ZFS system?

If the drive information is not getting passed to FreeNAS then the answer is no. How are you going to monitor drive life and errors if no SMART data is available to the system?

 

[StephenFry]

Any SMART info troubles have been solved, I wasn't clear about that. Some issues with my test motherboard and HBAs.

I can monitor the drives.

I realize that passive backplanes are much preferred over active ones; it's just that I ended up with the ones I did...

It's not even clear to me *if* these are active backplanes. Just one post on all of the internet says anything about this, and that doesn't seem enough data to go by...

Decent specs from AIC are nowhere to be found. Do active backplanes look different from passive ones? Different parts, etc? There's seemingly not a lot going on, but who knows.
This is it:

 

[Ericloewe]

I'm not sure what this active/passive talk is supposed to be about, but let's examine this:

• The backplane is actually four identical modules for four drives, with no expander. That means the drives are connected basically directly upstream.
• The little ICs in the area of the second drive are probably driving the front panel drive LEDs and stuff like that. Good luck getting them to do anything useful without the software stack they generally use

So it should be fine.

 

[StephenFry]

Very interesting, thank you!

I'm going to be testing this backplane and then decide if I like using it. Airflow in the box doesn't seem to be that great; I might use direct connections instead of this "wall" just because of that.
SuperMicro has some excellent cables for this, that are quite thin and easy to manipulate/bundle nicely. But that is a different topic.